We welcome Coombs and Trestman’s shift from “brain-only” models to a framework integrating body, motor, brain, sensory, and behavioral traits to explain the evolution of complex cognition (CC). However, we argue that cognitive complexification is inherently context dependent and cannot be meaningfully compared across taxa without considering their various niches and environments.

As cognitive complexification depends on ecological and environmental contexts, it is unlikely that a single, universal “complexification” of cognition exists. Whereas the authors successfully identified various interrelated factors leading to one form of CC, their approach may suffer from anthropocentric biases (Buckner, Reference Buckner2013). It selectively pre-defines CC based on human-inspired traits, like interrelated complexifications of bodily, behavioral, and sensorimotor capacities (e.g., high-resolution vision and high-speed locomotion). Hence, while this proposal may apply to humans and animals with similar evolutionary trajectories, it inadequately accounts for the full spectrum of CC. Below, we specifically identify three key problems of the framework.

Relevant abilities for CC were determined by tests tailored to human utility. An ecologically valid approach to cognition requires attending to both the environment and ecologies of the behaviors that organisms are expressing (Blitzer & Huebner, Reference Blitzer and Huebner2012). It must also acknowledge that species exploit their bodily and environmental resources in idiosyncratic species- and task-specific ways to achieve particular goals (Barrett, Reference Barrett2011). Yet, several cognitive skills identified in the target article, such as Theory of Mind (ToM) or mental time travel, primarily reflect human cognitive challenges. This implies a form of methodological and evaluative anthropocentrism forcing other species to become “aquatic or feathered apes” (Barrett, Reference Barrett2016; see also Buckner, Reference Buckner2013).

Experiments often introduce biases toward human physiologies and sensory modalities. Many animal cognition studies based on human cognitive abilities implicitly favor species with similar physiology, sensory modalities, and compatibility with experimental designs tailored to humans (Brebner et al., Reference Brebner, Loconsole, Hanley and Vasas2024; Caves, Nowicki, & Johnsen, Reference Caves, Nowicki and Johnsen2019).

For anatomy and physiology, tool use is a pertinent example. It heavily favors animals with appendages, as tools are commonly defined as handheld devices that aid in accomplishing a task (Merriam Webster, 2025) and are less often understood functionally as an external object to accomplish a particular task (Tariel-Adam et al., Reference Tariel-Adam, Toledo, O’Brien, Floeter and Brown2025). Thus, most tool use experiments feature organisms with flexible limbs or using their mouth in place of limbs (Tariel-Adam et al., Reference Tariel-Adam, Toledo, O’Brien, Floeter and Brown2025; Wynne, Reference Wynne and Udell2020). Only a minority of studies has even identified nontraditional tool use (Deb, Modak, & Balakrishnan, Reference Deb, Modak and Balakrishnan2020) or more exotic potential cases, such as net construction by spiders (Japyassú & Laland, Reference Japyassú and Laland2017), pit construction by ant larvae (Griffiths, Reference Griffiths1980), or even (as a non-animal example) the use of ants as tools by the “zombie fungus” Ophiocordyceps unilateralis.

In terms of sensory modalities, there is a clear anthropocentric bias toward vision-based testing. For instance, one of the highest “markers” of cognition, self-recognition, is frequently assessed via the mirror test (Bekoff, Allen, & Burghardt, Reference Bekoff, Allen and Burghardt2002). Yet, there have been few attempts to adjust this test for other sensory modalities (Bates et al., Reference Bates, Sayialel, Njiraini, Poole, Moss and Byrne2008; Horowitz, Reference Horowitz2017) such as olfaction, echolocation, or even just non-trichromatic, non-high-resolution vision. This results in the underestimation of CC in species lacking human vision. The author’s assumption that visual systems uniquely impose high processing demands is contestable (Gibson, Reference Gibson1979) and, moreover, it is unclear why senses like echolocation would be less cognitively demanding (Yong, Reference Yong2023). While the authors do not exclude the importance of other senses, it should be asked whether the traits recognized in their framework truly account for the natural diversity of sensory ecologies. The target article assumes that because animals that pass human-inspired tests show similar perceptual and physiological modalities, those must be CC traits. Yet, as those tests are not sufficiently adjusted to other ecologies, the only thing the authors identified is one path to CC.

Embodied frameworks should include ecological, social, and cultural factors. Most real-world thinking occurs in very particular environments (Anderson, Reference Anderson2003) that are shaped by ecological niches and social contexts. Thus, a key thesis of embodied cognition is that cognition does not just occur in the brain but in an integrated environment–organism complex (Barrett, Reference Barrett2011; Gibson, Reference Gibson1979). Yet, one may presume that the authors have ignored those factors because they don’t play an important role from an evolutionary perspective. This is a little bit misguided, though. Given that the ecological niche and social environment have always played a role in explaining evolutionary trait developments (Pocheville, Reference Pocheville, Heams, Huneman, Lecointre and Silberstein2015), it is surprising that the framework does not include such factors. Additionally, since the introduction of the concept of niche construction (Odling-Smee et al., Reference Odling-Smee, Laland and Feldman2003), it is even more clear that evolutionary trait development is not just a unidirectional process. Organisms actively shape their environments, affecting the possibilities of future transformations over evolutionary times (Laubichler & Renn, Reference Laubichler and Renn2015) and creating feedback loops in the ecological (Laland, Matthews, & Feldman, Reference Laland, Matthews and Feldman2016), sociocultural (Odling-Smee, Laland, & Feldman, Reference Odling-Smee, Laland and Feldman2003), and cognitive spaces (Bertolotti & Magnani, Reference Bertolotti and Magnani2017). Each evolutionary trajectory is then embedded in the organism’s ecology and the problem spaces of associated cognitive challenges only arise within that very ecology. Thus, ecology and social environments very much belong to the evolutionary story of organisms and cognitive systems.

Our critique should not be read as a rejection of the possibility of developing any trait-based comparative framework for the evolution of cognition. On the contrary, we applaud the authors for identifying one viable set of integrated body, sensory, and motor traits. We argue, however, that it is possible to develop other embodied trait complexes once we recognize how sensory-motor modalities, ecologies, and (social) environments relate and interact with each other. Alternatively, greater emphasis could be placed on what the authors identified as “cognitive and behavioral correlates”. While a number of those traits still have to be adjusted for organism ecologies, they are far more inclusive across contexts. Indeed, perhaps any investigation of cognition should begin with identifying the species-specific cognitive challenges and developing applicable traits from there. This may complicate the development of a multitrait framework, but by doing so we just may end up with a more inclusive approach to cognition that does not put the cart before the horse.